About me

I am a Ph.D. student at Cornell University, advised by Prof. Kevin Ellis. I completed my bachelor’s and master’s degrees at Shanghai Jiao Tong University.

My research interest lies in the generalizability of AI models, in particular, generalizing/adapting to out-of-distribution samples (e.g., inputs of larger scales, unseen combinations, and new concepts/domains). I currently focus on neuro-symbolic program synthesis to integrate neural networks with programmatic prior for better generalizability and interpretability. It involves fields such as program synthesis, latent abstraction learning, and learning to optimize.

Publications

• LLM-Guided Probabilistic Program Induction for POMDP Model Estimation
  Aidan Curtis, Hao Tang, Thiago Veloso, Kevin Ellis, Joshua Tenenbaum, Tom´as Lozano-P´erez, Leslie Pack Kaelbling
  CoRL 2025 [arxiv].
• Programmatic Video Prediction Using Large Language Models
  Hao Tang, Kevin Ellis, Suhas Lohit, Michael J Jones, Moitreya Chatterjee
  ICLR Workshop World Models 2025 [arxiv].
• Learning Abstract World Models with Neuro-Symbolic Predicates for Robot Planning
  Yichao Liang, Nishanth Kumar, Hao Tang, Adrian Weller, Joshua B Tenenbaum, Tom Silver, Jo˜ao F Henriques, Kevin Ellis
  ICLR 2025 [arxiv].
• Combining Induction and Transduction for Abstract Reasoning
  Wen-Ding Li*, Keya Hu*, Carter Larsen, Yuqing Wu, Simon Alford, Caleb Woo, Spencer M. Dunn, Hao Tang, Michelangelo Naim, Dat Nguyen, Wei-Long Zheng, Zenna Tavares, Yewen Pu†, Kevin Ellis†
  ICLR 2025 & Best Paper at ARCPrize [arxiv].
• WorldCoder, a Model-Based LLM Agent: Building World Models by Writing Code and Interacting with the Environment
  Hao Tang, Darren Key, and Kevin Ellis
  NeurIPS 2024 [project] [arxiv] [code].
• Code Repair with LLMs gives an Exploration-Exploitation Tradeoff
  Hao Tang, Keya Hu, Jin Peng Zhou, Sicheng Zhong, Wei-Long Zheng, Xujie Si, and Kevin Ellis
  NeurIPS 2024 [project] [arxiv] [code].
• From Perception to Programs: Regularize, Overparameterize, and Amortize
  Hao Tang, and Kevin Ellis
  ICML 2023 [arxiv],
  ICML Differentiable Everthing workshop 2023, PLDI MAPS symposium 2022.
• Towards Scale-Invariant Graph-related Problem Solving by Iterative Homogeneous GNNs
  Hao Tang, Zhiao Huang, Jiayuan Gu, Bao-Liang Lu, and Hao Su
  NeurIPS 2020 [arxiv] [code] [short-video] [poster] [pdf] [appendix].
• Refactoring Policy for Compositional Generalizability using Self-Supervised Object Proposals
  Tongzhou Mu*, Jiayuan Gu*, Zhiwei Jia, Hao Tang, and Hao Su
  NeurIPS 2020 [arxiv] [code].
• Belief Propagation Neural Networks
  Jonathan Kuck, Shuvam Chakraborty, Hao Tang, Rachel Luo, Jiaming Song, Ashish Sabharwal, and Stefano Ermon
  NeurIPS 2020 [arxiv].
• Emotion Recognition using Multimodal Residual LSTM Network
  Jiaxin Ma*, Hao Tang*, Wei-Long Zheng, and Bao-Liang Lu
  ACM Multimedia 2019 [pdf].
• Investigating Sex Differences in Classification of Five Emotions from EEG and Eye Movement Signals
  Lan-Qing Bao, Jie-Lin Qiu, Hao Tang, Wei-Long Zheng, and Bao-Liang Lu
  IEEE International Engineering in Medicine and Biology Conference (EMBC) 2019.
• Multimodal Emotion Recognition Using Deep Neural Networks
  Hao Tang, Wei Liu, Wei-Long Zheng, and Bao-Liang Lu
  International Conference on Neural Information Processing (ICONIP) 2017 [pdf].

Preprints

• PoE-World: Compositional World Modeling with Products of Programmatic Experts
  Wasu Top Piriyakulkij, Yichao Liang, Hao Tang, Adrian Weller, Marta Kryven, Kevin Ellis
  [arxiv].

Misc.

• Top Reviewer of NeurIPS, 2022.